Data integration
Zhiyuan Hu
07/01/2020
1 Preprocessing data
sceset is the SingleCellExperiment object that contains all the single cell data
sceset <- readRDS("../../scFT-paper_rds/20191016data_integration_all_sceset.rds")
dim(sceset)## [1] 22110 4330## Add secretory annotations
secretory <- readRDS("../../man_analysis2_20180629/rds/20190120Fresh_secretory_9clusters_clincluster.rds")
secretory$cell_type[secretory$clincluster_final == "C10"] <- "Cell cycle (C9)"
secretory$cell_type[secretory$clincluster_final == "C3"] <- "Differentiated (C3)"
secretory$cell_type[secretory$clincluster_final == "C4"] <- "KRT17 (C4)"
secretory$cell_type[secretory$clincluster_final == "C6"] <- "Stress (C6)"
secretory$cell_type[secretory$clincluster_final == "C8"] <- "EMT (C7)"
secretory$cell_type[secretory$clincluster_final == "C9"] <- "Immune (C8)"
secretory$cell_type[secretory$clincluster_fina %in% c("C1","C2","C5")] <- "Queiscent"
secretory$secretory_subtype <- secretory$cell_type
cancer_secretory <- secretory ## the clustering results of cancer_secretory
rm(secretory)1.1 Removing doublets
sceset$secretory_subtype <- cancer_secretory$cell_type[match(colnames(sceset), colnames(cancer_secretory))]
seurat <- as.Seurat(sceset)## Warning: Feature names cannot have underscores ('_'), replacing with dashes
## ('-')
## Warning: Feature names cannot have underscores ('_'), replacing with dashes
## ('-')seurat <- NormalizeData(seurat)
seurat <- ScaleData(object = seurat)
seurat <- FindVariableFeatures(seurat, selection.method = "vst", nfeatures = 2000)
seurat <- RunPCA(seurat)
seurat <- RunUMAP(seurat, dims = 1:15, seed.use = 12345)## Warning: Cannot add objects with duplicate keys (offending key: UMAP_),
## setting key to 'umap_'seurat <- RunTSNE(seurat, dims = 1:15, seed.use = 12345)
seurat <- FindNeighbors(seurat, dims = 1:15)
seurat <- FindClusters(seurat, resolution = 0.6)
# DimPlot(seurat, reduction = "tsne", label = T)
seurat$disease <- NA
seurat$disease[seurat$Patient2 %in% c(11543, 11545, 11553)] <- "ovarian"
seurat$disease[seurat$Patient2 %in% c(15066, 15091, 15072)] <- "endometrial"
seurat$disease[seurat$Patient2 %in% c(33572, 33778, 34350, 34659,35773)] <- "benign"
sweep.res.list_kidney <- paramSweep_v3(seurat, PCs = 1:15, sct = FALSE)
sweep.stats_kidney <- summarizeSweep(sweep.res.list_kidney, GT = FALSE)
bcmvn_kidney <- find.pK(sweep.stats_kidney)
## Homotypic Doublet Proportion Estimate
annotations <- seurat$seurat_clusters ## Clustering results
homotypic.prop <- modelHomotypic(annotations)
nExp_poi <- round(0.01*length(seurat@active.ident)) ## 1% doublet rate
nExp_poi.adj <- round(nExp_poi*(1-homotypic.prop))
## Run DoubletFinder with varying classification stringencies
seurat <- doubletFinder_v3(seurat, PCs = 1:10, pN = 0.25, pK = 0.005, nExp = nExp_poi, reuse.pANN = FALSE)
seurat <- doubletFinder_v3(seurat, PCs = 1:10, pN = 0.25, pK = 0.005, nExp = nExp_poi.adj, reuse.pANN = "pANN_0.25_0.005_43")
seurat@meta.data$DF_hi.lo <- seurat@meta.data$DF.classifications_0.25_0.005_43
seurat@meta.data$DF_hi.lo[which(seurat@meta.data$DF_hi.lo == "Doublet" & seurat@meta.data$DF.classifications_0.25_0.005_36 == "Singlet")] <- "Doublet_lo"
seurat@meta.data$DF_hi.lo[which(seurat@meta.data$DF_hi.lo == "Doublet")] <- "Doublet_hi"
# DimPlot(seurat, reduction = "tsne", group.by="DF_hi.lo", plot.order=c("Doublet_hi","Doublet_lo","Singlet"), colors.use=c("black","gold","red"))
colData(sceset)$DF_hi.lo <- seurat$DF_hi.lo
# all(colnames(seurat)==colnames(sceset))
benign_ft <- seurat[,seurat$DF_hi.lo == "Singlet" & seurat$type=="Benign"]
# saveRDS(benign_ft, "rds/20191030benignFT_1857cells_SeuratObj.rds")In total, 1857 cells from benign FT were left for analysis.
dim(benign_ft)## [1] 22110 18571.2 UMAP plots
# p1 <- DimPlot(benign_ft, reduction = "TSNE", group.by = "Patient2", pt = 0.4) +
# ggtitle(paste("t-SNE plot \n Benign FT (n = ", ncol(benign_ft), ")", sep = ""))
DimPlot(benign_ft, reduction = "UMAP", group.by = "Patient2", pt = 0.4) +
ggtitle(paste("UMAP plot \n Benign FT (n = ", ncol(benign_ft), ")", sep = ""))
# cowplot::plot_grid(p1,p2)
# ggsave("../revision_plots/SI/FigS3_benignFT_UMAPtSNE_colByPatient.png", width = 10, height = 4)FeaturePlot(benign_ft, reduction = "UMAP",features = c("PTPRC","COL1A1","EPCAM","KRT7","PAX8","CAPS"), ncol = 3, pt.size = 0.4)
# ggsave("../revision_plots/SI/FigS3_benignFT_UMAPcolourByMarkers.png", width = 14, height = 8)2 Filter out non-secretory cells
From the cells from benign donors we filtered out non-secretory cells.
# sum(benign_ft$seurat_clusters %in% c(8,10:12) | benign_ft$FACS == "CD45+EPCAM-") # 122 leukocytes
secretory <- benign_ft[,!benign_ft$seurat_clusters %in% c(8,10:12) & benign_ft$FACS != "CD45+EPCAM-"] # remove CD45+ cells
# sum(secretory$FACS == "STROMAL") # 11 stroma cells
secretory <- secretory[, !secretory$FACS %in% c("STROMAL")]
dim(secretory)## [1] 22110 17241724 FTE cells in total
# sum(secretory$seurat_clusters[secretory$type== "Benign"] %in% c(3, 7)) # 515 ciliated cells
secretory <- secretory[, !secretory$seurat_clusters %in% c(3, 7)] # remove cilicated cells
cpm <- log2(calculateCPM(secretory@assays$RNA@counts) + 1)
secretory <- secretory[, (cpm["KRT7",] > 0.5 | cpm["PAX8",] > 0.5) &
cpm["EPCAM",] > 0.5 &
cpm["PTPRC",] == 0
]
rm(cpm)
table(secretory$Patient2)##
## 33572 33778 34350 34659 35773
## 210 53 130 356 16# 33572 33778 34350 34659 35773
# 210 53 130 356 16
secretory_new <- secretory # secretory from the benign donors
# saveRDS(secretory, "rds/20191030benign_765secretory_seuratObj.rds")cancer_secretory <- as.Seurat(cancer_secretory)## Warning: Feature names cannot have underscores ('_'), replacing with dashes
## ('-')
## Warning: Feature names cannot have underscores ('_'), replacing with dashes
## ('-')cancer_secretory$type <- "Cancer"
secretory <- merge(secretory_new, cancer_secretory)
secretory$secretory_subtype[is.na(secretory$secretory_subtype) & secretory$type == "Cancer"] <- "Unclassified"
# saveRDS(secretory, "rds/seurat_secretory_merged20191007.rds")3 Batch effects
Before we integrated the secretory cells from cancer and benign patients, let’s first see the batch effects between them.
secretory <- NormalizeData(secretory)
secretory <- ScaleData(object = secretory)
secretory <- FindVariableFeatures(secretory, selection.method = "vst", nfeatures = 2000)
secretory <- RunPCA(secretory)
ElbowPlot(secretory, 40)
secretory <- RunUMAP(secretory, dims = 1:14, seed.use = 123345)
secretory <- RunTSNE(secretory, dims = 1:14, seed.use = 123345)
secretory <- FindNeighbors(secretory, dims = 1:14, seed.use = 123345)
secretory <- FindClusters(secretory, resolution = 0.6)It is clear that batch effects exist between the two batches as shown in the figure below. So the next step we did is to remove the batch effects.
p1 <- DimPlot(secretory, reduction = "umap", label = T)## Warning: Using `as.character()` on a quosure is deprecated as of rlang 0.3.0.
## Please use `as_label()` or `as_name()` instead.
## This warning is displayed once per session.p2 <- DimPlot(secretory, reduction = "umap", group.by = "type")
cowplot::plot_grid(p1,p2)
3.1 Data integration (standard workflow)
This section integrated different batches.
seu_list <- list()
seu_list[[1]] <- secretory[,secretory$type == "Cancer"]
seu_list[[1]] <- seu_list[[1]][,seu_list[[1]]$secretory_subtype != "Immune (C8)" & seu_list[[1]]$secretory_subtype != "Unclassified"]
seu_list[[2]] <- secretory[,secretory$Patient2 %in% c("34350","33778","35773","33572")]
seu_list[[3]] <- secretory[,secretory$Patient2 == "34659"]
for(itor in 1:3){
seu_list[[itor]] <- NormalizeData(seu_list[[itor]], verbose = FALSE)
seu_list[[itor]] <- FindVariableFeatures(seu_list[[itor]],
selection.method = "vst",
nfeatures = 1000, verbose = FALSE)
}
n_dim <- 15
ftsc.anchors <- FindIntegrationAnchors(object.list = seu_list,
dims = 1:n_dim,
k.anchor = 5) ## Computing 2000 integration features## Scaling features for provided objects## Finding all pairwise anchors## Running CCA## Merging objects## Finding neighborhoods## Finding anchors## Found 1779 anchors## Filtering anchors## Retained 1167 anchors## Extracting within-dataset neighbors## Running CCA## Merging objects## Finding neighborhoods## Finding anchors## Found 1645 anchors## Filtering anchors## Retained 1085 anchors## Extracting within-dataset neighbors## Running CCA## Merging objects## Finding neighborhoods## Finding anchors## Found 1454 anchors## Filtering anchors## Retained 1156 anchors## Extracting within-dataset neighborsftsc.integrated <- IntegrateData(anchorset = ftsc.anchors, dims = 1:n_dim)## Merging dataset 3 into 2## Extracting anchors for merged samples## Finding integration vectors## Finding integration vector weights## Integrating data## Merging dataset 1 into 2 3## Extracting anchors for merged samples## Finding integration vectors## Finding integration vector weights## Integrating dataDefaultAssay(ftsc.integrated) <- "integrated"
# Run the standard workflow for visualization and clustering
ftsc.integrated <- ScaleData(ftsc.integrated, verbose = FALSE)
ftsc.integrated <- RunPCA(ftsc.integrated, npcs = n_dim, verbose = FALSE)
ftsc.integrated <- RunUMAP(ftsc.integrated, reduction = "pca", dims = 1:n_dim, seed.use = 12345)p1 <- DimPlot(ftsc.integrated, reduction = "umap", group.by = "type", pt.size = 0.5)
p2 <- DimPlot(ftsc.integrated,
reduction = "umap",
group.by = "secretory_subtype", pt.size = 0.5)
cowplot::plot_grid(p1,p2, rel_widths = c(1,1.1))
p1 <- DimPlot(ftsc.integrated[,ftsc.integrated$type == "Cancer" &
!is.na(ftsc.integrated$secretory_subtype)],
reduction = "umap", group.by = "secretory_subtype") + ggtitle("Cancer")
p2 <- DimPlot(ftsc.integrated[,ftsc.integrated$type == "Benign"],
group.by = "type", reduction = "umap") + ggtitle("Benign")
cowplot::plot_grid(p1,p2)
3.2 Clustering results
ftsc.integrated <- FindNeighbors(ftsc.integrated, dims = 1:n_dim, seed.use = 12345, k.param = 5)## Computing nearest neighbor graph## Computing SNNftsc.integrated <- FindClusters(ftsc.integrated, resolution = 0.6)## Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck
##
## Number of nodes: 1453
## Number of edges: 18343
##
## Running Louvain algorithm...
## Maximum modularity in 10 random starts: 0.7525
## Number of communities: 7
## Elapsed time: 0 seconds# ftsc.integrated <- readRDS("../../scFT-paper_rds/ftsc.integrated_20191013.rds")
DimPlot(ftsc.integrated, label = T, reduction = "umap")
table(ftsc.integrated$secretory_subtype[ftsc.integrated$type == "Cancer"], ftsc.integrated$seurat_clusters[ftsc.integrated$type == "Cancer"])##
## 0 1 2 3 4 5 6
## Cell cycle (C9) 10 2 1 3 0 0 7
## Differentiated (C3) 21 1 65 80 0 0 0
## EMT (C7) 3 2 0 0 1 34 0
## KRT17 (C4) 14 10 1 20 139 2 0
## Stress (C6) 161 22 72 11 4 0 2ftsc.integrated$secretory_subtype_benign <- NA
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Benign" & ftsc.integrated$seurat_clusters %in% c(4)] <- "KRT17"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Benign" & ftsc.integrated$seurat_clusters %in% c(9)] <- "Cell cycle"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Benign" & ftsc.integrated$seurat_clusters %in% c(5)] <- "EMT"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Benign" & ftsc.integrated$seurat_clusters %in% c(2,3)] <- "Differentiated"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Cancer" & ftsc.integrated$secretory_subtype == "KRT17 (C4)"] <- "KRT17"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Cancer" & ftsc.integrated$secretory_subtype == "Cell cycle (C9)"] <- "Cell cycle"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Cancer" & ftsc.integrated$secretory_subtype == "EMT (C7)"] <- "EMT"
ftsc.integrated$secretory_subtype_benign[ftsc.integrated$type == "Cancer" & ftsc.integrated$secretory_subtype == "Differentiated (C3)"] <- "Differentiated"3.3 Scoring the cells from benign samples
new_markers <- readRDS("../../scFT-paper_rds/20190213new_markers.rds")
ftsc.integrated$KRT17_score <- colSums(ftsc.integrated@assays$RNA@data[new_markers$C4,])
ftsc.integrated$CC_score <- colSums(ftsc.integrated@assays$RNA@data[new_markers$C10,])
ftsc.integrated$EMT_score <- colSums(ftsc.integrated@assays$RNA@data[new_markers$EMT,])
ftsc.integrated$Diff_score <- colSums(ftsc.integrated@assays$RNA@data[new_markers$C3,])ftsc.sceset <- as.SingleCellExperiment(ftsc.integrated)
df.plot <- cbind(ftsc.sceset@reducedDims$UMAP,
ftsc.sceset@colData[,c("KRT17_score", "CC_score", "EMT_score", "Diff_score", "type", "secretory_subtype_benign")])
df.plot <- as.data.frame(df.plot)
# df.plot <- df.plot[!is.na(df.plot$secretory_subtype_benign),]
df.plot1 <- df.plot[df.plot$type == "Benign",]
df.plot2 <- df.plot[df.plot$type == "Cancer",]
scaleScore <- function(x, max.limit = 3) {
x <- scale(x, center = T, scale = T)
x[x > max.limit] <- max.limit
return(x)
}
df.plot1$KRT17_score <- scaleScore(df.plot1$KRT17_score)
df.plot1$CC_score <- scaleScore(df.plot1$CC_score)
df.plot1$Diff_score <- scaleScore(df.plot1$Diff_score)
df.plot1$EMT_score <- scaleScore(df.plot1$EMT_score)
df.plot2$KRT17_score <- scaleScore(df.plot2$KRT17_score)
df.plot2$CC_score <- scaleScore(df.plot2$CC_score)
df.plot2$Diff_score <- scaleScore(df.plot2$Diff_score)
df.plot2$EMT_score <- scaleScore(df.plot2$EMT_score)
plotScoreUmap <- function(dat, column){
ggplot(data=dat, aes_string(x="UMAP_1", y="UMAP_2", fill = column)) +
geom_point(alpha = 0.5, pch = 21, col = "grey80") +
theme_classic() +
scale_fill_gradient2(midpoint=0,
low="grey90", mid="white",high="red", space ="Lab" ) +
theme(plot.title = element_text(size=14, face="bold", hjust = 0.5))
}
plist <- list()
plist[[1]] <- plotScoreUmap(dat = df.plot1, column = "KRT17_score") + ggtitle("Benign (KRT17 score)")
plist[[2]] <- plotScoreUmap(dat = df.plot1, column = "CC_score") + ggtitle("Benign (cell cycle score)")
plist[[3]] <- plotScoreUmap(dat = df.plot1, column = "Diff_score") + ggtitle("Benign (differentiated score)")
plist[[4]] <- plotScoreUmap(dat = df.plot1, column = "EMT_score") + ggtitle("Benign (EMT score)")
plist[[5]] <- plotScoreUmap(dat = df.plot2, column = "KRT17_score") + ggtitle("Cancer (KRT17 score)")
plist[[6]] <- plotScoreUmap(dat = df.plot2, column = "CC_score") + ggtitle("Cancer (cell cycle score)")
plist[[7]] <- plotScoreUmap(dat = df.plot2, column = "Diff_score") + ggtitle("Cancer (differentiated score)")
plist[[8]] <- plotScoreUmap(dat = df.plot2, column = "EMT_score") + ggtitle("Cancer (EMT score)")
cowplot::plot_grid(plotlist = plist, ncol = 4)
# ggsave("../revision_plots/SI/benign_cancer_subtype_score_plots.png", height = 8, width = 18)
# ggsave("../revision_plots/SI/benign_cancer_subtype_score_plots.tiff", height = 8, width = 18)
# ggsave("../revision_plots/SI/benign_cancer_subtype_score_plots.pdf", height = 8, width = 18)The expression of marker genes is expressed in the genes at the similar location.
# saveRDS(ftsc.integrated, "rds/ftsc.integrated_20191013.rds")4 Technical
sessionInfo()## R version 3.5.2 (2018-12-20)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS Mojave 10.14.4
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRlapack.dylib
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] grid parallel stats4 stats graphics grDevices utils
## [8] datasets methods base
##
## other attached packages:
## [1] fields_9.6 maps_3.3.0
## [3] spam_2.2-2 dotCall64_1.0-0
## [5] modes_0.7.0 ROCR_1.0-7
## [7] gplots_3.0.1.1 KernSmooth_2.23-15
## [9] DoubletFinder_2.0.1 reshape2_1.4.3
## [11] scales_1.0.0 Seurat_3.0.2
## [13] dplyr_0.7.8 edgeR_3.24.3
## [15] limma_3.38.3 scater_1.10.1
## [17] ggplot2_3.2.0.9000 SingleCellExperiment_1.4.1
## [19] SummarizedExperiment_1.12.0 DelayedArray_0.8.0
## [21] BiocParallel_1.16.5 matrixStats_0.54.0
## [23] Biobase_2.42.0 GenomicRanges_1.34.0
## [25] GenomeInfoDb_1.18.1 IRanges_2.16.0
## [27] S4Vectors_0.20.1 BiocGenerics_0.28.0
##
## loaded via a namespace (and not attached):
## [1] Rtsne_0.15 ggbeeswarm_0.6.0
## [3] colorspace_1.4-0 ggridges_0.5.1
## [5] XVector_0.22.0 listenv_0.7.0
## [7] npsurv_0.4-0 ggrepel_0.8.0
## [9] codetools_0.2-16 splines_3.5.2
## [11] R.methodsS3_1.7.1 lsei_1.2-0
## [13] knitr_1.21 jsonlite_1.6
## [15] ica_1.0-2 cluster_2.0.7-1
## [17] png_0.1-7 R.oo_1.22.0
## [19] sctransform_0.2.0 HDF5Array_1.10.1
## [21] compiler_3.5.2 httr_1.4.0
## [23] assertthat_0.2.0 Matrix_1.2-15
## [25] lazyeval_0.2.1 htmltools_0.3.6
## [27] tools_3.5.2 rsvd_1.0.2
## [29] bindrcpp_0.2.2 igraph_1.2.3
## [31] gtable_0.2.0 glue_1.3.0
## [33] GenomeInfoDbData_1.2.0 RANN_2.6.1
## [35] Rcpp_1.0.0 gdata_2.18.0
## [37] ape_5.2 nlme_3.1-137
## [39] DelayedMatrixStats_1.4.0 gbRd_0.4-11
## [41] lmtest_0.9-36 xfun_0.4
## [43] stringr_1.4.0 globals_0.12.4
## [45] irlba_2.3.3 gtools_3.8.1
## [47] future_1.14.0 zlibbioc_1.28.0
## [49] MASS_7.3-51.1 zoo_1.8-4
## [51] rhdf5_2.26.2 RColorBrewer_1.1-2
## [53] yaml_2.2.0 reticulate_1.10
## [55] pbapply_1.4-0 gridExtra_2.3
## [57] stringi_1.2.4 caTools_1.17.1.1
## [59] bibtex_0.4.2 Rdpack_0.10-1
## [61] SDMTools_1.1-221 rlang_0.4.0
## [63] pkgconfig_2.0.2 bitops_1.0-6
## [65] evaluate_0.13 lattice_0.20-38
## [67] purrr_0.3.0 Rhdf5lib_1.4.2
## [69] bindr_0.1.1 labeling_0.3
## [71] htmlwidgets_1.3 cowplot_0.9.4
## [73] tidyselect_0.2.5 plyr_1.8.4
## [75] magrittr_1.5 R6_2.3.0
## [77] pillar_1.3.1 withr_2.1.2
## [79] fitdistrplus_1.0-14 survival_2.43-3
## [81] RCurl_1.95-4.11 tsne_0.1-3
## [83] tibble_2.0.1 future.apply_1.3.0
## [85] crayon_1.3.4 plotly_4.8.0
## [87] rmarkdown_1.11 viridis_0.5.1
## [89] locfit_1.5-9.1 data.table_1.12.0
## [91] metap_1.1 digest_0.6.18
## [93] tidyr_0.8.2 R.utils_2.7.0
## [95] munsell_0.5.0 beeswarm_0.2.3
## [97] viridisLite_0.3.0 vipor_0.4.5# R version 3.5.2 (2018-12-20)
# Platform: x86_64-apple-darwin15.6.0 (64-bit)
# Running under: macOS Mojave 10.14.4
#
# Matrix products: default
# BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
# LAPACK: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRlapack.dylib
#
# locale:
# [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
#
# attached base packages:
# [1] parallel stats4 stats graphics grDevices utils datasets methods base
#
# other attached packages:
# [1] class_7.3-15 DoubletFinder_2.0.1 reshape2_1.4.3 scales_1.0.0
# [5] Seurat_3.0.2 dplyr_0.7.8 edgeR_3.24.3 limma_3.38.3
# [9] scater_1.10.1 ggplot2_3.2.0.9000 SingleCellExperiment_1.4.1 SummarizedExperiment_1.12.0
# [13] DelayedArray_0.8.0 BiocParallel_1.16.5 matrixStats_0.54.0 Biobase_2.42.0
# [17] GenomicRanges_1.34.0 GenomeInfoDb_1.18.1 IRanges_2.16.0 S4Vectors_0.20.1
# [21] BiocGenerics_0.28.0
#
# loaded via a namespace (and not attached):
# [1] backports_1.1.3 plyr_1.8.4 igraph_1.2.3 lazyeval_0.2.1
# [5] splines_3.5.2 listenv_0.7.0 usethis_1.4.0 digest_0.6.18
# [9] htmltools_0.3.6 viridis_0.5.1 gdata_2.18.0 magrittr_1.5
# [13] memoise_1.1.0 cluster_2.0.7-1 ROCR_1.0-7 remotes_2.0.2
# [17] globals_0.12.4 RcppParallel_4.4.3 R.utils_2.7.0 prettyunits_1.0.2
# [21] colorspace_1.4-0 ggrepel_0.8.0 xfun_0.4 callr_3.1.1
# [25] crayon_1.3.4 RCurl_1.95-4.11 jsonlite_1.6 bindr_0.1.1
# [29] survival_2.43-3 zoo_1.8-4 ape_5.2 glue_1.3.0
# [33] gtable_0.2.0 zlibbioc_1.28.0 XVector_0.22.0 leiden_0.3.1
# [37] pkgbuild_1.0.2 Rhdf5lib_1.4.2 future.apply_1.3.0 HDF5Array_1.10.1
# [41] bibtex_0.4.2 Rcpp_1.0.0 metap_1.1 viridisLite_0.3.0
# [45] reticulate_1.10 rsvd_1.0.2 SDMTools_1.1-221 tsne_0.1-3
# [49] htmlwidgets_1.3 httr_1.4.0 gplots_3.0.1.1 RColorBrewer_1.1-2
# [53] ica_1.0-2 pkgconfig_2.0.2 R.methodsS3_1.7.1 uwot_0.1.3
# [57] locfit_1.5-9.1 tidyselect_0.2.5 labeling_0.3 rlang_0.4.0
# [61] munsell_0.5.0 tools_3.5.2 cli_1.0.1 devtools_2.0.1
# [65] ggridges_0.5.1 stringr_1.4.0 yaml_2.2.0 npsurv_0.4-0
# [69] fs_1.2.6 processx_3.2.1 knitr_1.21 fitdistrplus_1.0-14
# [73] caTools_1.17.1.1 purrr_0.3.0 RANN_2.6.1 bindrcpp_0.2.2
# [77] pbapply_1.4-0 future_1.14.0 nlme_3.1-137 R.oo_1.22.0
# [81] compiler_3.5.2 rstudioapi_0.9.0 curl_3.3 beeswarm_0.2.3
# [85] plotly_4.8.0 png_0.1-7 lsei_1.2-0 tibble_2.0.1
# [89] stringi_1.2.4 ps_1.3.0 desc_1.2.0 RSpectra_0.13-1
# [93] lattice_0.20-38 Matrix_1.2-15 pillar_1.3.1 Rdpack_0.10-1
# [97] lmtest_0.9-36 RcppAnnoy_0.0.11 data.table_1.12.0 cowplot_0.9.4
# [101] bitops_1.0-6 irlba_2.3.3 gbRd_0.4-11 R6_2.3.0
# [105] KernSmooth_2.23-15 gridExtra_2.3 vipor_0.4.5 sessioninfo_1.1.1
# [109] codetools_0.2-16 MASS_7.3-51.1 gtools_3.8.1 assertthat_0.2.0
# [113] pkgload_1.0.2 rhdf5_2.26.2 rprojroot_1.3-2 withr_2.1.2
# [117] sctransform_0.2.0 GenomeInfoDbData_1.2.0 grid_3.5.2 tidyr_0.8.2
# [121] DelayedMatrixStats_1.4.0 Rtsne_0.15 ggbeeswarm_